the impacts of water infrastructure and climate change on ......infrastructure and climate change on...
TRANSCRIPT
Luna Bharati, Guillaume Lacombe, Pabitra Gurung, Priyantha Jayakody, Chu Thai Hoanh and Vladimir Smakhtin
The Impacts of Water Infrastructure and Climate Change on the Hydrology of the Upper Ganges River Basin
142
RESEARCHIWMI
R E P O R T
I n t e r n a t i o n a lWater ManagementI n s t i t u t e
Research Reports
The publications in this series cover a wide range of subjects—from computer
modeling to experience with water user associations—and vary in content from
directly applicable research to more basic studies, on which applied work ultimately
depends. Some research reports are narrowly focused, analytical and detailed
empirical studies; others are wide-ranging and synthetic overviews of generic
problems.
Although most of the reports are published by IWMI staff and their
collaborators, we welcome contributions from others. Each report is reviewed
internally by IWMI staff, and by external reviewers. The reports are published and
distributed both in hard copy and electronically (www.iwmi.org) and where possible
���� ����� ���� ������������ �� � ������� ��� ������� ����������� ������������
may be copied freely and cited with due acknowledgment.
About IWMI
IWMI’s mission is to improve the management of land and water resources for food, livelihoods and the environment. In serving this mission, IWMI concentrates
on the integration of policies, technologies and management systems to achieve
������� ���������� ��� ���� ������������������� �� ���� ������� ��� ��� ���� ���
irrigation and water and land resources.
i
International Water Management Institute
P O Box 2075, Colombo, Sri Lanka
IWMI Research Report 142
The Impacts of Water Infrastructure and
Climate Change on the Hydrology of the
Upper Ganges River Basin
Luna Bharati, Guillaume Lacombe, Pabitra Gurung, Priyantha Jayakody, Chu Thai Hoanh and Vladimir Smakhtin
The authors: Luna Bharati is a Senior Researcher and Head of the Nepal office of the International
Water Management Institute (IWMI) in Kathmandu, Nepal ([email protected]); Guillaume Lacombe is
a Researcher based at the Laos office of IWMI in Vientiane, Lao PDR ([email protected]); Pabitra
Gurung is a Research Officer at the Nepal office of IWMI in Kathmandu, Nepal ([email protected]);
Priyantha Jayakody was a Research Officer based at the headquarters of IWMI in Colombo, Sri Lanka,
and is now a PhD student at the Mississippi State University, Agriculture and Biological Department in
Starkville, Mississippi, United States ([email protected]); Chu Thai Hoanh is a Principal Researcher
and Acting Head of the Laos office of IWMI in Vientiane, Lao PDR ([email protected]); and Vladimir
Smakhtin is Theme Leader - Water Availability and Access at the headquarters of IWMI in Colombo, Sri
Lanka ([email protected]).
Bharati, L.; Lacombe, G.; Gurung, P.; Jayakody, P.; Hoanh, C. T.; Smakhtin, V. 2011. The impacts of water infrastructure and climate change on the hydrology of the Upper Ganges River Basin. Colombo, Sri Lanka:
International Water Management Institute. 36p. (IWMI Research Report 142). doi: 10.5337/2011.210
/ water resources / river basins / climate change / hydrology / simulation models / precipitation / evapotranspiration / statistical methods / water balance / water yield / irrigation water / India / Ganges River /
ISSN 1026-0862
ISBN 978-92-9090-744-2
Copyright © 2011, by IWMI. All rights reserved. IWMI encourages the use of its material provided that the
organization is acknowledged and kept informed in all such instances.
Cover picture shows Ganga River upstream of Rishikesh, Uttarakhand, India (Photo credit: Vladimir
Smakhtin, IWMI)
Please send inquiries and comments to: [email protected]
A free copy of this publication can be downloaded at
www.iwmi.org/Publications/IWMI_Research_Reports/index.aspx
Acknowledgements
The authors would like to thank the Indian Institute of Tropical Meteorology (IITM) for providing regional
climate change projection data (Providing REgional Climates for Impacts Studies (PRECIS)) for our analysis.
We would also like to thank Nitin Kaushal, Senior Manager, World Wide Fund for Nature (WWF)-India, for
helping us with data acquisition and WWF-India for funding this study.
Project
This research study was initiated as part of WWF-India’s ‘Living Ganga Programme’.
Collaborators
The following organizations collaborated in the research conducted for this report.
International Water Management Institute (IWMI)
World Wide Fund for Nature (WWF)-India
Donors
Funding for this research was provided by the following organization.
World Wide Fund for Nature (WWF)-India
Water ManagementI n t e r n a t i o n a l
I n s t i t u t e
v
v
Contents
Summary vii
Introduction 1
Methods and Data 4
Results and Discussion 14
Conclusions 25
References 27
vi
vii
and mountainous upper areas of the UGB. On
an annual average, present-day flows throughout
the UGB are about 2-8% lower than in naturalized
conditions. The percentage of flow reduction is the
highest during the dry months as water is being
withdrawn for irrigation. Dry and wet season flows
under CC scenario A2 (scenario corresponding
to high population growth with slower per capita
economic growth and technological change) are
lower than those in present climate conditions
at upstream locations, but higher at downstream
locations of the UGB. Flows under CC scenario B2
(corresponding to moderate population growth and
economic development with less rapid and more
diverse technological change) are systematically
higher and lower than those under CC scenario
A2 during dry and wet seasons, respectively. The
dates of minimum daily discharges are highly
variable among stations and between different CC
scenarios, while the dates of maximum flow are
delayed downstream as a result of the delay in
the onset of the monsoon in the lower parts of the
basin. The report also provides actual simulated
discharge time series data for all simulated
scenarios, in the overall attempt to augment the
river flow data for this important river basin and to
facilitate the use of these data by any interested
party.
Summary
Provision of detailed continuous long-term
hydrological time series data for any river basin is
critical for estimation, planning and management
of its current and future water resources. Most of
the river basins in India are data poor, including
its iconic river – Ganga (Ganges). This study
assessed the variability of flows under present
and ‘naturalized’ basin conditions in the Upper
Ganges Basin (UGB) (area of over 87,000 square
kilometers (km2)). The naturalized basin conditions
are those that existed prior to the development of
multiple water regulation structures, and hence
may be seen as a reference condition, a starting
point, against which to evaluate the impacts
of planned basin development, as well as the
impacts of future climate change (CC) on basin
water resources. The later impacts are also part
of the study: the PRECIS regional climate model
(RCM) was used to generate climate projections
for the UGB, with subsequent simulations of
future river flows. Results show that the annual
average precipitation, actual evapotranspiration
(ET) and net water yields of the whole basin
were 1,192 millimeters (mm), 416 mm and 615
mm, respectively. However, there were large
variations in both temporal and spatial distribution
of these components. Precipitation, ET and water
yields were found to be higher in the forested
1
The Impacts of Water Infrastructure and Climate Change on the Hydrology of the Upper Ganges River Basin
Luna Bharati, Guillaume Lacombe, Pabitra Gurung, Priyantha Jayakody, Chu Thai Hoanh and Vladimir Smakhtin
Introduction
The Ganges River System originates in the
Central Himalayas, and extends into the alluvial
Gangetic Plains and drains into the Indian Ocean
at the Bay of Bengal. Its basin area (1.09 million
km2) spreads across India (79%), Nepal (13%),
Bangladesh (4%) and China (4%). The river is
of high importance to riparian countries with an
estimated 410 million people directly or indirectly
depending on it (Verghese and Iyer 1993). In
the upstream mountainous regions, hydropower
is the main focus of development with mega
and micro projects either under construction or
being planned in both Nepal and India. After
the main river channel reaches the plains, it
is highly regulated with dams, barrages and
associated irrigation canals. All this infrastructure
development and abstractions affects the river’s
flow regime and reduces flows, which, in turn,
impacts downstream water availability, water
quality and riverine ecosystems. Furthermore,
there are concerns that CC is likely to exacerbate
the water scarcity problem in the Ganges Basin.
Therefore, modeling the hydrology of the basin is
critical for estimation, planning and management
of current and future water resources.
To operate a hydrological model, reliable data
on climatological variables such as temperature,
precipitation, evaporation, etc., over space and
time are necessary. For analysis of the past, such
information can be derived from observational
data sets. However, for assessment of the future
with the possible impacts of CC, the hydrological
models can be driven with the output from the
general circulation model (GCM) (IPCC 1996;
Akhtar et al. 2009). However, the resolutions of
GCMs are currently constrained by computational
and physical reasons to 200 kilometers (km)
for climate change predictions and are too
course for hydrological modeling at basin scale.
In order to increase the spatial resolution of
these predictions, one method that is used is
statistical downscaling techniques, which have
been developed in the last decades (e.g., Wilby
et al. 1999; Bergström et al. 2001; Pilling and
Jones 2002; Guo et al. 2002; Arnell 2003; Booij
2005; Benestad et al. 2008). A second option
is the use of dynamical downscaling (e.g., Hay
et al. 2002; Hay and Clark 2003; Fowler and
Kilsby 2007; Leander and Buishand 2007).
Dynamical downscaling fits output from GCMs
into regional meteorological models. It involves
using numerical meteorological modeling to
reflect how global patterns affect local weather
conditions. The high horizontal resolution of a
RCM (about 10-50 km) is more appropriate for
resolving the small-scale features of topography
and land use. Furthermore, the high resolution
of RCM is ideal to capture the spatial variability
of precipitation as input into hydrological models
(Gutowski et al. 2003; Akhtar et al. 2009),
and provide better representation of mountain
areas affected by the amount of rainfall and the
location of windward rainy areas and downwind
rain shadow areas (Jones et al. 2004).
2
Hydrological simulations using RCM output in
data-sparse basins such as the Ganges involves
several problems, including uncertainties in inputs,
model parameters and model structure (Akhtar
et al. 2009). The main disadvantages of RCMs
are that they inherit the large-scale errors of their
driving GCM model and require large amounts of
boundary data previously archived from relevant
GCM experiments. Additional uncertainties can be
linked to the local scale patterns in downscaling of
temperature, precipitation and evapotranspiration
in a specific basin (Bergström et al. 2001; Guo et
al. 2002; Akhtar et al. 2009).
Although modeling studies that analyzed
the impacts of water infrastructure development
on the hydrology and water resources in other
parts of the world are available, studies focusing
on the Ganges Basin are limited. The National
Communication (NATCOM) project by the Ministry
of Environment and Forests, Government of
India, quantified the impact of CC on water
resources of all major Indian river systems
(Gosain et al. 2006). This study used the Hadley
Centre Regional Climate Model 2 (HadRM2) daily
weather data as input to run the Soil and Water
Assessment Tool (SWAT) hydrological model to
determine the spatio-temporal water availability
in the river systems and to calculate basin water
balances. This study suggests that precipitation,
evapotranspiration and runoff will increase by
approximately 10% in the Ganges Basin. The
NATCOM study, however, did not consider the
effect of water infrastructure development in the
basin and modeled the Ganges without water
abstraction and use. Furthermore, the simulations
were not validated against observed flow data,
making the model results uncertain.
The Water Evaluation And Planning (WEAP)
model was used by de Condappa (2009)
for large-scale assessment of surface water
resources in the Indus and Ganges basins,
with a special focus on the contribution of snow
and glaciers to flow. Several, relatively simple
scenarios of changes in glaciated area were
simulated. Results suggest that glaciers play the
role of buffers against interannual variability in
precipitation, in particular, during years with a
weak monsoon. However, the impacts of CC and
water use in the basin were not fully assessed.
Some other modeling studies examined, in detail,
the hydrological regime of individual glaciers
in the UGB (Singh et al. 2008), rather than the
impacts of water use and CC on basin-wide water
resources. Similarly, Seidel et al. (2000) modeled
the runoff regime of the Ganges and Brahmaputra
basins, accounting for precipitation, remotely
sensed snow covered areas and temperatures
using the Snowmelt Runoff Model (SRM). They
found that the already high risk of floods during
the period July to September is slightly increased
with CC. Numerous papers can be found on
the impact of water resources development and
CC on downstream areas of Bangladesh (e.g.,
Ahmad et al. 2001; Jian et al. 2009; Mirza 2004;
Rahaman 2009). These studies, however, do not
assess the dynamics of water availability and use
in upstream areas within Nepal and India.
Apart from the general paucity of hydrological
and water resources studies with fine spatial
resolution in the Ganges Basin, the inherent
problem of this basin is the availabil ity of
observed discharge data, against which models
can be calibrated and validated. Discharge data in
the Himalayan part of the basin are scarce due to
lack of measurement stations. In the downstream
plains, although discharge data from gauging
stations exist, these data are not accessible to
the public due to national security laws in India.
This leaves most of the hydrology studies of the
Ganges, which are carried out by the government
agencies, being classified and not accessible in
the public domain. In addition, simulated data are
also not widely shared, hence impeding their use
in subsequent water resource applications.
The UGB (upstream of Kanpur Barrage;
ca tchment a rea 87 ,787 km2) (F igure 1)
encompasses various physiographic conditions,
is of great cultural and spiritual importance for
the country, yet, it is already highly regulated with
multiple water structures (with more plans on the
way), and will most likely experience significant
changes in hydrology (with significant economic
and social implications) due to CC.
This study is part of WWF-India’s “Living
Ganga Programme.” The main objective of this
programme is to develop and promote approaches
3
for sustainable water resources management,
including environmental flows which conserve
biodiversity and support livelihoods under present
and changing climate scenarios (WWF 2011).
In the present study, the flow variability of the
catchment is characterized for four scenarios
corresponding to pairwise combinations of present
and future water infrastructure development and
climate conditions, as detailed in the section,
Methods and Data. Results are presented in two
parts: 1) water balances for subcatchments for
present and ‘naturalized’ conditions (prior to the
development of multiple regulation structures), and
2) several indicators of hydrological variability that
characterize the likely impacts of CC at both present
and ‘naturalized’ flow regimes. For purposes of this
study, naturalized flows are defined as ‘free flowing
flows in the mainstream without dams and barrages,
and irrigation diversions’. These assessments
are the prerequisites for the subsequent detailed
water allocation modeling under present and future
flow regimes in the UGB. In addition, the aim of
the simulation modeling here is to provide freely
available flow time series for the UGB, in order to
enhance data availability and initiate hydrological
data sharing. All simulated data are also available
via IWMI’s water data portal (http://waterdata.iwmi.
org/).
FIGURE 1. Upper Ganges Basin (UGB) with locations of barrages, reservoirs, hydrometeorological stations and ‘environmental
flow (EF) sites’, where environmental flow assessment was undertaken under the “Living Ganga Programme.”
4
Methods and Data
The analytical framework of the study is presented
in Figure 2 and detailed in the following sections.
The SWAT hydrological model is used to simulate
flows.
The steps of the analytical framework are
detailed below.
1. SWAT model setup for the UGB: a model
with 21 subbasins was setup using a Digital
Elevation Model (DEM), soil and land use/land
cover (LULC) maps and flow data.
2. Collection of observed climate data: daily
data for five climate variables (precipitation,
maximum and minimum temperature, solar
radiation, wind speed and relative humidity)
measured inside and around the basin from
1971 to 2005 were collected.
3. Calibration and validation of SWAT model against observed flow data over the period
2000-2005.
4. Selection of a RCM for simulation of future cl imate condit ions: output t ime series
simulated by PRECIS over the periods 1961-
1990 and 2071-2100 were provided by IITM.
5. Adjus tment o f c l ima te change da ta : PRECIS t ime ser ies data for the gr id
cells encompassing climatic stations were
Analysis of Model Results and Reporting
Running Model for Different Scenarios
(i) Present Scenario without CC (1971 - 2005)
(ii) Naturalized Scenario without CC (1971 - 2005)
(iii) Present Scenario with CC-A2 (2071 - 2100)
(iv) Present Scenario with CC-B2 (2071 - 2100)
Reset ModelAdjustment of CC Data
(Observed versus Baseline)
Regional PRECIS CC DataBL [1961 - 1990]
A2 [2071 - 2100]
B2 [2071 - 2100]
Model Calibration and Validation(2000 - 2005)
Model Setup
Spatial Maps (DEM, Soil and LULC)
Observed Climate Data (1971 - 2005) Precipitation
Temperature
Relative humidity
Wind speed
Solar radiation
Observed Flow Data
FIGURE 2. Analytical framework. Note: BL = baseline scenario; A2 = Climate conditions as projected by PRECIS under A2
scenarios; B2 = Climate conditions as projected by PRECIS under B2 scenarios
5
compared to actual records over the period
1961-1990 to statistically determine the
required bias adjustments. These adjustments
were then applied to the projection period
2071-2100 based on the assumption that
the same bias occurs in both simulation and
projection periods.
6. Setting study scenarios: based on the
objectives of the study, four scenarios with
different land and water use, and climate
conditions were established. These are
( i ) present condi t ion scenar io (water
infrastructure development as of 2005 and
present climate as measured from 1971 to
2005); (ii) ‘naturalized’ scenario, assuming
that no water infrastructure were built under
present climate conditions (1971-2005); (iii)
climate change scenarios, assuming that
water infrastructure is that of 2005, under
2071-2100 climate conditions as projected
by PRECIS under A2 scenarios; and (iv) B2
scenarios.
7. Scenario simulation using SWAT model.
8. Analysis of simulated results: simulated
water balances were compared between
present and ‘naturalized’ conditions. The
impacts of c l imate change and water
infrastructure developments on flow regimes
at the four ‘EF sites’ were characterized
using indicators of hydrological alterations.
This report presents the main results of this
analytical framework.
SWAT Model Description and Setup
SWAT is a process-based continuous hydrological
mode l tha t p red ic ts the impac t o f land
management practices on water, sediment and
agricultural chemical yields in complex basins
with varying soils, land use and management
conditions (Arnold et al. 1998; Srinivasan et al.
1998). The main components of the model include
climate, hydrology, erosion, soil temperature, plant
growth, nutrients, pesticides, land management,
channel and reservoir routing. SWAT divides a
basin into subbasins. Subbasins are connected
through a stream channel. Each subbasin is
further divided into Hydrological Response Units
(HRU). A HRU is a unique combination of soil
and vegetation types. SWAT simulates hydrology,
vegetation growth and management practices at
the HRU level.
The hydrological cycle is simulated by SWAT
using the following water balance equation (1):
)1(SWt = SWo + (Rday - Qsurf - Ea - wseep - Qgw )Σ
n
i = 1
where: SWt: Final soil water content (mm); SWo: Initial soil water content (mm); t: Time in days; Rday: pre-
cipitation on day i (mm); Qsurf : surface runoff on day i (mm); Ea: actual evapotranspiration on day i (mm);
wseep: percolation on day i (mm); and Qgw: return flow on day i (mm).
Basin subdivision allows differences in
evapotranspiration for various crops and soils
to be simulated. Runoff is predicted separately
for each subbasin and routed to obtain the total
runoff for the basin. This increases the accuracy
and gives a much better physical description of
the water balance. More detailed descriptions of
the model can be found in Arnold et al. (1998),
Srinivasan et al. (1998) and Neitsch et al. 2005.
The SWAT model was chosen for this study
because it can be used in large agricultural river
basin scales to also simulate crop water use.
As actual data for irrigation distribution was not
available, the calculated crop water use helped
determine the irrigation water requirements in the
basin. The SWAT model has been extensively
used in the USA and elsewhere for calculating
water balances of primarily agricultural catchments
(e.g., Jha 2011; Bharati and Jayakody 2011; Garg
et al. 2011).
6
The SWAT model requires three basic files for
delineating the basin into subbasins and HRUs:
Digital Elevation Model (DEM), soil map and Land
Use/Land Cover (LULC) map. Figure 3(a) shows
the DEM for the basin using 90 m Shuttle Radar
Topography Mission (SRTM) data.
The elevation in the UGB ranges from 100 m
in the lower plains to 7,500 m in the upper
mountain region. Some mountain peaks in the
headwater basin are permanently covered by
snow. Figure 3(b) shows the land use map which
was developed using the LandSat image from
2003. According to this map, around 65% of
the basin is occupied by agriculture. The main
crop types (identified from the land use map
and district statistics) are wheat, maize, rice,
sugarcane, pearl millet and potato. About 25%
of the area of the UGB is covered by forests,
mostly in the upper mountains. For the naturalized
scenarios, as there are no water provisions for
irrigation, the irrigated crops such as rice and
sugarcane are replaced by rainfed crops from the
region, i.e., lentils and wheat.
Figure 3(c) shows the soil map of the basin.
There are nine soil types in the UGB. Lithosols
dominate the upper, steep mountainous areas
and are very shallow and erodible. Cambisols and
luvisols are found in the lower areas. Cambisols
are developed in medium- and fine-textured
material derived from alluvial, colluvial and
aeolian deposits. Most of these soils make good
agricultural land. Luvisols are tropical soils mostly
used by farmers because of its ease of cultivation,
but they are greatly affected by water erosion and
loss in fertility. Annual average precipitation in the
UGB ranges from 550 to 2,500 mm (Figure 3(d)). A
major part of the rainfall is due to the southwestern
monsoon from July to October. Wet season
corresponds to the period July to November and
the dry season extends from December to June.
The UGB mainstream is highly regulated with
dams, barrages and corresponding canal systems
(Figure 1). The two main dams, Ramganga and
Tehri, became operational in 1988 and 2008, and
have total storage capacities of 2,448 and 3,540
106 m
3, respectively. There are three main canal
systems. The Upper Ganga canal takes off from
the right flank of the Bhimgoda barrage with a
head discharge of 190 m3/s, and irrigates an area
of 2 million hectares (Mha). The Madhya Ganga
canal provides annual irrigation water to 178,000
hectares (ha). The Lower Ganga canal from the
Narora weir irrigates 0.5 Mha.
During the winter season, a part of the runoff
in the basin is generated through contributions
from snowmelt and glacier melt. Therefore, for
this study, snowmelt was computed using the
SWAT model. In the model, when the mean
daily air temperature is less than the snowfall
temperature, the precipitation within a HRU
is classified as snow and the l iquid water
equivalent of the snow precipitation is added
to the snowpack. The snowpack increases with
additional snowfall, but decreases with snowmelt
or sublimation. In the model, snowmelt is
controlled by the air and snowpack temperature,
the melting rate and the areal coverage of snow.
Snowmelt is then included with rainfall in the
calculations of runoff and percolation. Further
information on snowmelt calculations can be
found in Neitsch et al. 2005. Wang and Melesse
(2005) evaluated the performance of snowmelt
hydrology of the SWAT model by simulating
streamflows for the Wild Rice River watershed
(located in the USA), and found that the SWAT
model had a good performance on simulating the
monthly, seasonal and annual mean discharges
and a satisfactory performance on predicting
the daily discharges. When analyzed alone, the
daily streamflows during the spring, which were
predominantly generated from melting snow,
could be predicted with an acceptable accuracy,
and the corresponding monthly and seasonal
mean discharges could be simulated very well.
The Gangotri glacier is located within the UGB.
Therefore, in order to include the contribution of
this glacier in the present model setup, discharge
data collected from a gauging site very close to
the snout of the glacier (Singh et al. 2006) was
added to the relevant subbasin.
7
FIGURE 3. The maps of the UGB used for modeling, with numbers and boundaries of subbasins used in hydrological simulations.
Note: FAO = Food and Agriculture Organization of the United Nations.
(a) Digital Elevation Model, SRTM
(c) Soil Map, FAO
(b) Land use map, Land Sat TM, 2003
(d) Mean Annual Precipitation (1971-2005)
8
Available Observed Time Series Data
The SWAT model requires time series of observed
climate data, including precipitation, minimum
and maximum temperature, sunshine duration,
wind speed and relative humidity. Table 1 lists the
climate stations used in the model. The locations
of stations are shown in Figure 1. The SWAT
model uses the data of a climate station nearest
to the centroid of each subcatchment as an input
for that subcatchment (Figure 3d).
Table 2 presents details of the flow stations
used for calibration and validation of the model.
Locations of the flow stations are shown in Figure
1. Due to restrictions on the distribution of Ganges
data from the Indian Central Water Commission
(CWC), only a very short time series of data at
some barrages were available. Although daily
observed flow data were available from Narora,
only monthly time series data were available for
the other sites. As the model works with daily
time steps, simulated daily flow values had to be
accumulated into monthly values for calibration
and validation. The existing dams, barrages and
irrigation deliveries were incorporated into the
model using available salient features from the
relevant barrage/dam authorities.
Station code1 Name Available record Available data type
R T S W H
42111 Dehradun2 1970-2005 x x x x x
42103 Ambala2 1970-2004 x x x x x
8207 Simla2 1989-2005 x x x x x
42140 Roorkee2 1970-1994;
2002-2005 x x x x x
42182 Delhi2 1970-2005 x x x x x
42366 Kanpur 1970-1974;
1986-1995 x x
42471 Fatehpur 1970-2005 x x
42189 Bareilly2 1970-2005 x x x x x
42260 Agra 1970-2005 x x
42262 Aligarh 1970-2005 x x
42143 Najibad2 1970-2005 x x x x x
42147 Mukteshwar2 1970-2005 x x x x x
42148 Pant Nagar2 1970-2005 x x x x x
42265 Mainpuri 1970-2005 x x x
42665 Shajapur 1970-2005 x
42266 Shahjahanpur 1970-2005 x x x x
Notes: 1 Station codes correspond to locations shown in Figure 1.
2 Station has large data gaps.
R = Precipitation; T = Minimum and maximum temperature; S = Sunshine duration; W = Wind speed; H = Relative humidity.
TABLE 1. Details of the meteorological data used.
9
����������� �������������� ���������������� ��� � ��������������� ���� ���������� ��� ���
Station code Location Catchment Available Type of data
area (km2) record
������� �� �!���� �"#$%$� �&� ���$$��'�������$$*� +������� ����� ������������!
Flow_2 Narora 29,840 January 2000 - June 2005 Daily spill release from the dam
Flow_3 Kanpur 87,790 June 2003 - December 2005 Monthly spill release from the dam
� � � � /��0� �!���������������
Model Calibration and Validation
The period from January 1, 1970 to December
31, 1971 was used to warm-up the model. The
available data between 2000 and 2005 from each
station were divided into two sets, each of them
including the same number of daily observations.
The first and second sets were used for model
calibration and validation, respectively. Model
parameters were calibrated simultaneously for all
three flow stations.
The model performance was determined by
calculating the coefficient of determination (R2)
and the Nash-Sutcliffe Efficiency (NSE) criterion.
R2 and NSE values for each simulation are
presented in Table 3. The model performance
over cal ibrat ion and val idation periods is
acceptable, according to the model performance
ratings proposed by Liu and De Smedt (2004).
In addition, comparisons were made between
the measured and simulated annual water flow
volumes. The differences in volume ranged from
22 to 25% and from 7 to 9% during calibration
and validation periods, respectively.
Figure 4 shows observed precipitation,
observed and simulated flows for inflow into the
Bhimgoda barrage, outflow from Narora barrage
and outflow from Kanpur barrage. Although
flows are regulated at these sites, observed and
simulated hydrographs match very well. This adds
confidence to the results that are presented in the
following sections.
Downscaling of Climate Model Data
Climate data from PRECIS were used as input
to the SWAT hydrological model in order to
assess future river flow scenarios. PRECIS
is an atmospheric and land surface model
developed at the Meteorological Office Hadley
Centre, UK, for generating high-resolution
climate change information for many regions of
the world (Jones et al. 2004). It has a spatial
resolution of 0.22° x 0.22°. Climate data from the
GCM, Hadley Centre Coupled Model, version 3
(HadCM3) (Jones et al. 2004), were downscaled
with PRECIS for the UGB under A2 and B2
Special Report on Emission Scenarios (SRES)
scenarios (IPCC 2000) by IITM. A2 and B2 are
two climate change SRES scenarios studied
by the Intergovernmental Panel on Climate
Change (IPCC). A2 corresponds to a story line
of high population growth with slower per capita
economic growth and technological change, and B2
corresponds to a story line of moderate population
growth and economic development with less rapid
and more diverse technological change. PRECIS
data used in that study were extracted from 15
grid cells corresponding to the location of the
15 meteorological stations previously described.
These time series data cover the periods 1961-
1990 and 2071-2100, and include four variables:
precipitation, temperature, wind speed and relative
humidity. Although these PRECIS data result from a
downscaled global climate model which accounts for
regional climate and topographic characteristics, they
still exhibit discrepancies with regard to observed
meteorological data. For instance, the mean
absolute relative difference in annual precipitation
depths between PRECIS and observed time series
over the baseline period 1970-1990 is about 55%.
Therefore, PRECIS data were adjusted in such a
way that, at each of the 15 station locations, the
main statistical properties of adjusted PRECIS
output (mean and standard deviation) match those
of the historical data. This statistical downscaling
approach is described in Bouwer et al. (2004).
10
TA
BLE
3. A
sum
mary
of th
e S
WA
T m
odel p
erf
orm
ance e
valu
ation.
���
��2
��
���
�����
��� �� ���3
���
�� ��4�
��
��
����0����
��
�����
����0
�
S
tatistic C
alib
ration
V
alid
ation
C
alib
ration
Valid
ation
Valu
e
Ratings
Valu
e
Ratings
Observ
ed
Sim
ula
ted
Diffe
rence
Observ
ed
Sim
ula
ted
Diffe
rence
Flo
w_1
April 1,
January
1,
R2
0.8
6
> 0
.85
0.8
4
(0.6
5 -
0.8
5)
1,1
52
1,4
37
24.7
%
1,0
17
1,0
90
-7.2
%
(Bhim
goda)
2002 -
2004 -
D
ecem
ber
D
ecem
ber
Excelle
nt
V
ery
good
31, 2003
31, 2005
N
SE
0.6
7
(0.6
5 -
0.8
5)
0.8
1
(0.6
5 -
0.8
5)
V
ery
good
V
ery
good
Flo
w_2
January
1,
Jan 1
, R
2
0.9
2
(> 0
.85)
0.8
8
(> 0
.85)
905
1,1
03
21.8
%
694
754
8.6
%
(Naro
ra)
2000 –
2003 -
D
ecem
ber
June
Excelle
nt
E
xcelle
nt
31, 2002
30, 2005
N
SE
0.9
0
(> 0
.85)
0.8
5
(0.6
5 -
0.8
5)
E
xcelle
nt
V
ery
good
Flo
w_3
June 1
, June 1
,
R2
0.8
6
(> 0
.85)
0.6
0
(0.6
0 -
0.6
4)
756
923
22.1
%
622
663
6.7
%
(Kanpur)
2003 -
2005 –
O
cto
ber
D
ecem
ber
Excelle
nt
G
ood
3
1,
20
03
3
1,
20
05
and
N
SE
0.8
2
(0.6
5 -
0.8
5)
0.6
6
(0.6
5 -
0.8
5)
June 1
,
V
ery
good
V
ery
good
2004 -
O
cto
ber
31,
2004
Sta
tion
co
de
an
d
location
Ca
libra
tio
n
period
Va
lida
tio
n
period
11
FIGURE 4. Observed and simulated flows at (a) Bhimgoda barrage, (b) Narora barrage, and (c) Kanpur barrage.
0
400
800
1,200
1,600
2,0000
2,000
4,000
6,000
8,000
10,000
Jan-03 May-03 Oct-03 Mar-04 Aug-04 Jan-05 Jun-05 Nov-05
Prec
ipit
atio
n (m
m)
Flo
w (m
3 /s)
(c)
Precipitation Simulated flow Observed flow
0
400
800
1,200
1,600
2,0000
2,000
4,000
6,000
8,000
10,000
Jan-00 Oct-00 Aug-01 Jun-02 Apr-03 Feb-04 Dec-04 Oct-05
Prec
ipit
atio
n (m
m)
Flo
w (m
3 /s)
(b)
Precipitation Simulated flow Observed flow
R2 = 0.92NSE = 0.90
R2 = 0.86NSE = 0.82 NSE = 0.66
Calibration Validation
R2 = 0.88NSE = 0.85
R2 = 0.60
Calibration Validation
R2 = 0.86NSE = 0.67
R2 = 0.84NSE = 0.81
0
400
800
1,200
1,600
2,0000
2,000
4,000
6,000
8,000
10,000
Jan-02 Jul-02 Jan-03 Jul-03 Jan-04 Jul-04 Jan-05 Jul-05 Jan-06
Prec
ipit
atio
n (m
m)
Flo
w (m
3 /s)
(a)
Precipitation Simulated flow Observed flow
Calibration Validation
(a)
(b)
(c)
12
The period 1961-2008 was used to calculate the
standard deviation, and average of PRECIS and
observed data. For each of the four downscaled
climate variables, specific adjustment rules were
defined and are detailed below.
Precipitation
It is assumed that, for each month of the year,
the proportion of dry days (daily precipitation
= 0 mm) should be identical in both data sets
(PRECIS and observation). In order to meet this
hypothesis, the distribution of non-rainy days in
the original PRECIS data set was modified before
it was statistically downscaled, as described in
Bouwer et al. (2004). The successive steps of the
calculations are described below.
i) The average proportion (P) of rainy days
(> 0 mm) for each month of the year was
determined for both PRECIS and observed
data sets for each meteorological station.
P was found to be significantly higher for
PRECIS data, i.e., PRECIS data include fewer
dry days than observations. This difference
is explained by the fact that observed data
are point-based while PRECIS data are grid-
based (each cell extends over an area of 0.22
x 0.22 square degrees - about 25 x 25 km2).
Thus, the probability of having a dry day over
this area is much lower than that of having a
dry day at a point location.
ii) For each month o f the year, a da i ly
precipi tat ion threshold H (mm/day) is
determined in such a way that P% of PRECIS
daily precipitation values are higher than H.
iii) For each month of the year and for both data
sets (all observed values and PRECIS values
higher than H), the means, MOBSERVATION and
MPRECIS, and standard deviations, STOBSERVATION
and STPRECIS, respectively, of rainy day depths
(> H) were calculated.
iv) For each month of the year, adjusted PRECIS
daily values ‘y’ were calculated from the
original PRECIS data ‘x’ as shown in Equation
(2) below.
Temperature
As the PRECIS data set does not include
maximum and minimum temperature but only
daily averages, data adjustment was based on
mean observed daily temperature calculated from
observed maximum and minimum values. For
each month of the year, average and standard
deviation of daily temperatures were calculated for
observed and PRECIS data using available daily
values for the period 1961-2008. Data adjustment
was performed similarly to precipitation data,
although the first step of precipitation adjustment
(threshold definition) was not necessary in the
case of temperature.
Furthermore, the SWAT model required
maximum and minimum temperature as an
input and the adjusted PRECIS temperatures
were in daily average (xAVG). Therefore, the
observed daily data were used to calculate
monthly averages and standard deviations of
maximum, minimum and average temperature
(MOBS, MAX, MOBS, MIN, MOBS, AVG, STOBS, MAX, STOBS,
MIN, and STOBS, AVG), and then the following
equations were used to calculate the maximum
daily PRECIS temperature (TMAX) (Equation (3))
and minimum daily PRECIS temperatures (TMIN)
(Equation (4)).
)2(If x > H, then y = ( x - MPRECIS ) + MOBSERVATION
If x < H, then y = 0
STOBSERVATIONSTPRECIS
13
)3(Maximum daily PRECIS temperature, T MAX = ( x AVG - MOBS,AVG) + MOBS,MAX
STOBS, MAX
STOBS, AVG
)4(Minimum daily PRECIS temperature, T MIN = ( x AVG - MOBS,AVG) + MOBS,MIN
STOBS, MIN
STOBS, AVG
Wind speed
In the data set of observations, wind speed values
are measured 2 meters above the ground surface.
In contrast, PRECIS data correspond to wind
speeds 10 meters above the ground surface. This
difference in elevations results in higher PRECIS
wind velocities, as wind is slowed down by
frictional resistance close to the surface. As this
difference is observed for each month of the year,
a unique reduction coefficient (0.38) was applied
to all PRECIS daily values. This coefficient was
calculated by dividing the mean observed daily
wind speed value averaged over the period 1961-
2008 by that obtained from PRECIS data. Further
adjustment steps consisted of applying Equation
(1) to PRECIS wind speed daily values, similarly
to other climate variables.
Relative humidity
The adjustment of relative humidity is similar to
that of temperature and includes the same steps.
A further adjustment, when required, consisted
of replacing values exceeding 100% by ‘100%’.
Such cases (less than 1% of adjusted values),
was due to the following reasons: in some cases,
the range of observed relative humidity values
over the baseline period was found to exceed
that of the PRECIS data. This resulted in higher
monthly averages and/or standard deviations for
observed values, in comparison with those of
PRECIS data. The latter, when used for the data
adjustment, could produce daily relative humidity
values exceeding 100%.
Scenario Setting
Four scenarios were simulated. In scenario 1,
the water infrastructure system up until the year
2005 was included. This scenario includes water
abstractions from dams and barrages. The Tehri
Dam (Figure 1), which became operational in
2008, is not included in this scenario. Irrigated
crops such as rice, wheat, corn, finger millet,
sugarcane and potato represent the major crop
types during present conditions. Scenario 2
represents the ‘naturalized’ condition without any
artificial flow regulation. Simulated flows for this
scenario were produced after having removed
all water infrastructures from the calibrated
model. In this scenario, farming areas were
characterized by rainfed crops such as mung
bean and wheat. Most of the non-agricultural
land is covered by natural forest. Scenarios 3
and 4 correspond to the present conditions of
water infrastructure development, abstraction
(up until 2005) with CC scenario adjusted from
PRECIS time series data under SRES A2 and
B2, respectively. Table 4 provides a detailed
description of the scenarios.
TABLE 4. Description of simulated scenarios.
No. Water infrastructure development Climate input data
1 Present (as of 2005) Observed data 1971-2005
2 Naturalized conditions Observed data 1971-2005
3 Present (as of 2005) Adjusted PRECIS data over period 2071-2100 under A2 scenario
4 Present (as of 2005) Adjusted PRECIS data over period 2071-2100 under B2 scenario
14
Results and Discussion
Water Balance Under Present Conditions
and Natural Conditions (Scenarios 1 and 2)
Figure 5 shows the annual average water balance
for all 21 subbasins of the UGB, and Figure
6 shows the mean monthly water balance of
the whole basin, both under Scenario 1. Four
hydrological components are considered, i.e.,
precipitation, actual evapotranspiration (ET), net
water yield (which is a routed runoff from the
subbasin) and balance closure. The term ‘balance
closure’ includes groundwater recharge, change
in soil moisture storage in the vadose zone and
model inaccuracies.
Annual average precipitation, actual ET and
net water yield of the whole basin were 1,192,
416 and 615 mm, respectively. However, there
was a large variation in the spatial distribution of
these components. Precipitation, ET and water
yield were found to be higher in the forested and
mountainous upper areas. In the upper subbasins,
water yield is higher than ET. However, in some
of the lower subbasins dominated by agriculture,
ET values were higher than water yield. Water
balances from the lower part of the catchment,
containing irrigated areas, are affected by water
regulation through barrages, dams and canals.
The large network of canals is transferring water
from one subbasin to irrigate crops in another
subbasin. Therefore, it is difficult to determine
the accuracy of the runoff calculations from each
subbasin. However, the ET and precipitation
figures are useful in characterizing the water
availability and use in each of the subbasins. The
maximum precipitation of 2,504 mm occurred in
subbasin 3 and minimum precipitation of 536 mm
occurred in subbasins 8 and 11 (see also Figure
3(d)). Furthermore, the maximum ET of 671 mm
occurred in subbasin 7 and the minimum ET of
177 mm occurred in subbasin 10.
The mean monthly results from 1971 to 2005
(Figure 6) show that there are large temporal
variations in the water balance components. The
maximum precipitation of 338 mm occurred during
August and a minimum of 7 mm occurred in
November. Similarly, water yields are also much
higher during the monsoon months as compared to
the dry season. ET, however, which is more related to
land cover, was found to be lowest during the winter
months, i.e., November-January (post-rice harvest).
-3,000
-1,500
0
1,500
3,000
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
An
nu
al a
vera
ge
wat
er b
alan
ce(m
m)
Subbasin number
Precipitation Actual ET Net water yield Balance closure
FIGURE 5. Annual average water balance results of model simulation at the subbasin level (1971-2005). The subbasin
numbers are given in Figure 3.
15
Simulation of Natural Flow Conditions for
the Four EF Sites (Scenario 2)
Simulating non-regulated, pre-development
flow regimes is important to determine the
re ference hydro logy, against which f low
changes in a basin can be measured. Strictly
speaking, this constitutes the assessment
of water actually available for all uses and
development – a starting point in sustainable
basin planning and management. Natural flow
simulations are also required to assess EF
– the flow regimes that are required for the
ecological health of the river. Results from
the EF assessment for the UGB are beyond
the scope of this paper and are discussed in
another report (WWF 2011).
Natural flows for four locations (sites EF1-
EF4, coordinates in Table 5) are presented in the
following sections. Although natural flows have
been simulated for the whole basin, these four
locations were chosen for this study as they are
representative of different agroecological zones
in the river stretch used for this study. These
locations were also sites for the EF assessment
study and are, therefore, referred to as EF sites
in this report.
Simulated daily flow data at the four EF sites
under scenarios 1 and 2 were summed up to
monthly and annual time steps, and are presented
in the tables and figures below. As already
mentioned above, the modeling period only went
up to 2005, so the effect of the Tehri Dam, which
became operational in 2008, was not considered.
Therefore, for site EF1, only natural flows have
been reported because water is neither stored in
dams nor abstracted for agriculture upstream of
this site.
-400
-200
0
200
400
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Mea
n m
on
thly
wat
er b
alan
ce(m
m)
Month
Precipitation Actual ET Net water yield Balance closure
FIGURE 6. Mean monthly water balance results of model simulation (1971-2005) for the entire UGB - at Kanpur.
TABLE 5. Location and names of representative sites along the UGB (see also Figure 1).
Site code Name Latitude Longitude
EF1 Kaudiyala Rishikesh 30°04’29” N 78°30’09” E
EF2 Narora 29°22'22” N 78°2'20” E
EF3 Kachla Bridge 27°55’59” N 78°51’42” E
EF4 Bithur (Kanpur) 26°36'59” N 80°16'29” E
16
Figure 7 shows the plots of annual water
flow volume and monthly water flow volume at
site EF4, Bithur (Kanpur), which is located near
the outlet of the UGB. Table 6 shows simulation
flow results for the four EF sites including
simulated present f lows for EF sites 2-4.
Comparison between natural and present flows
showed that, on average, the present annual
water volume is 7%, 2% and 8% lower than
in the natural conditions at EF sites 2, 3 and
4 (Narora, Kachla Bridge and Bithur (Kanpur)),
respectively. At all sites, the percentage of flow
reduction is highest during the dry months as
water is being withdrawn for irrigation. At Narora
(site EF2), maximum flow reduction is 70% in
February. Similarly, at site EF3, the maximum
flow reduction is 35% in February and at site
EF4, the maximum flow reduction is 58% also
in February (Table 6). In April, and some other
months, mean flow volume of the present
conditions exceeds the flow volume of the natural
conditions. This is caused by the basin flow
transfer from Ramganga Dam. Although high
percentages of flow reduction occur in the dry
season, the contribution of flow in this season to
annual flow is low (less than 10%).
FIGURE 7. (a) Annual flow totals, and (b) average monthly flow distribution for Bithur (Kanpur).
0
2,000
4,000
6,000
8,000
10,000
12,000
14,000
16,000
18,000
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Vo
lum
e (M
CM
)
Month
EF4 - Bithur (Kanpur)
Natural monthly flow volume Present monthly flow volume
(a)
0
20,000
40,000
60,000
80,000
100,000
1971
1973
1975
1977
1979
1981
1983
1985
1987
1989
1991
1993
1995
1997
1999
2001
2003
2005
Vo
lum
e (M
CM
)
Year
EF4 - Bithur (Kanpur)
Annual natural flow volume Annual present flow volume
(b)
(a)
(b)
17
EF
sites
Annual
sim
ula
ted f
low
A
nnual
Perc
enta
ge o
f m
onth
ly r
eduction i
n f
low
s (
%)
v
olu
me (
MC
M)
re
duction
in f
low
s
(volu
me
N
atu
ral
P
resent
(%))
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
Dec
EF
1
(Kaudiy
ala
/
Ris
hik
esh)
38,4
45
38,4
45*
0 (
0%
) 0
0
0
0
0
0
0
0
0
0
0
0
EF
2
(Naro
ra)
42,6
08
39,6
32
2,9
76 (
7%
) 68
70
44
-10
30
24
0
0
2
4
19
18
EF
3
(Kachla
Bridge)
42,9
09
42,1
65
774 (
2%
) 27
35
14
-60
7
16
-1
-1
0
2
12
18
EF
4
(Bithur
(Kanpur)
) 57,0
61
52,2
68
4,7
93 (
8%
) 44
58
35
-44
12
31
11
7
5
2
4
10
TA
BLE
6. S
imula
ted flo
w r
esults a
t E
F s
ites.
Not
e: *
No r
educt
ion c
om
pare
d w
ith n
atu
ral c
onditi
ons,
beca
use
there
was
no w
ate
r use
upst
ream
of th
is s
ite d
uring the s
tudy
period.
18
Figure 8 shows the comparison of water
yield (runoff from subbasin) distribution at the
subbasin level for the natural condition as well
as the present condition. As can be expected,
water yield or runoff, from the upper forested
subbasins, have not changed. However, there
are reduced flows during the present condition
from the lower subbasins, mainly due to water
withdrawals for agricultural production. There are
a few subbasins in the lower catchment where
water yields have increased and this is a result
of water transfers from the Upper and Madhya
Ganga canals.
Figure 9 shows flow duration curves (FDC)
in present and natural conditions at the EF sites.
These curves indicate that flows are lower in
the present condition in comparison with natural
conditions.
FIGURE 8. Percentage change in mean annual net water yields at present condition in comparison to natural condition of model
simulation at subbasin level (i.e., present-natural).
19
FIGURE 9. Flow duration curves (FDC) for (a) Kaudiyala/Rishikesh, (b) Narora, (c) Kachla Bridge, and (d) Bithur (Kanpur) sites.
100
1,000
10,000
0 20 40 60 80 100
Flo
w (m
3 /s)
Flo
w (m
3 /s)
Percentage of exceedence (%)
EF1 - Kaudiyala/Rishikesh
10
100
1,000
10,000
0 20 40 60 80 100
Percentage of exceedence (%)
EF2 - Narora
Natural Present
100
1,000
10,000
0 20 40 60 80 100
Flo
w (m
3 /s)
Flo
w (m
3 /s)
Percentage of exceedence (%)
EF3 - Kachla Bridge
Natural Present
10
100
1,000
10,000
0 20 40 60 80 100
Percentage of exceedence (%)
EF4 - Bithur (Kanpur)
Natural Present
(a) (b)
(c) (d)
Table 7 shows the difference in flow rates
between simulations of natural and present
conditions for 40% to 90% of percentile of
exceedence. Usually, water-related infrastructure
such as hydropower plants or irrigation systems
are designed taking into consideration a design
discharge corresponding to 40% to 90%. The
difference in flows affects any future planning
of water resources development as well as
allocations for environmental flows in the river.
TABLE 7. Difference in flows between simulations of the natural and present conditions at EF sites.
������ �������3���0���'&����4�3�3/s)
Percentage of exceedence EF2 (Narora) EF3 (Kachla Bridge) EF4 (Bithur (Kanpur))
40% 232 99 156
50% 120 28 115
60% 93 17 48
70% 89 25 54
80% 114 41 68
90% 126 37 88
20
�������������� ������������������
Characteristics Under Different Simulation
Scenarios Including CC (Scenarios 1-4)
Five indicators of hydrological variability were
derived from simulated time series: mean daily dry
season flows, mean daily wet season flows, the
date of maximum flow, the date of minimum flow
and the number of reversals reflecting the rate of
change in river discharge.
Hydrological Variability and Alteration Caused by Dams Under Present Climate Conditions
The analysis first focuses on flows simulated from
observed precipitation, either under the present
conditions of water infrastructure development
or in naturalized conditions. Figure 10 displays
mean dry and wet season discharges computed
from mean monthly discharge values. In natural
conditions, dry season flows remain stable all
along the river stream, with values ranging from
302 to 340 m3/s. This flow regime indicates that
river flow mostly originates from the melting of
glacier and snow cover while the flow contribution
from the water table drainage is negligible. During
the wet season, mean daily discharge consistently
increases from upstream to downstream, reflecting
the successive flow contributions of the tributaries,
and collecting surface runoff produced by
monsoon rainstorms. While the impact of water
infrastructure remains moderate during the wet
season (flows under present condition are 6%
lower than flows produced in natural conditions),
relative flow changes during the dry season are
of a higher magnitude, especially at EF sites 2,
3 and 4, as no dam exists in the EF1 headwater
catchment. Between the sites EF1 and EF4, the
dams have induced a 25% flow decrease as a
result of dry season irrigation.
Figure 11 displays the mean Julian date of
1-day minimum and maximum flow at four EF
sites, under natural and present conditions. In
natural conditions, 1-day minimum and maximum
flow conditions are delayed downstream. This
shift in the flow regime results from the difference
in the onset of the monsoon between upstream
and downstream parts of the basin. First rains
of the wet season occur in the upper part
of the basin (Figure 12). Under the present
conditions, the downstream shift in the date
of the 1-day maximum flow is similar to that
observed under natural conditions. In contrast,
the 1-day minimum flow does not follow this
pattern. Dates at sites EF1 and EF3 are similar
(dates = 22) while the 1-day minimum flow
occurs much earlier at sites EF2 and EF4 (date
= 1 for EF2; date = 2 for EF4). This alteration of
the natural flow regime results from the operation
of the dams located upstream of the sites EF2
and EF4, storing flows at the beginning of the
year. The impact of these dams on the date of
the 1-day maximum flow is imperceptible, as the
range of flow variations caused by the operation
of the dams is much lower than the mean river
discharge during this period of the year.
FIGURE 10. Mean dry and wet season daily discharge for the period 1971-2005 at four EF sites, under naturalized (natural)
and present conditions.
0
50
100
150
200
250
300
350
400
EF1 EF2 EF3 EF4
Mea
n d
isch
arg
e (m
3 /s)
Mea
n d
isch
arg
e (m
3 /s)
Dry season
0
500
1,000
1,500
2,000
2,500
3,000
3,500
4,000
EF1 EF2 EF3 EF4
Wet season
Natural
Present
21
FIGURE 11. Timing of annual extreme water conditions.
0
20
40
60
80
100
120
EF1 EF2 EF3 EF4
Julia
n d
ate
Date of minimum flow
220
222
224
226
228
230
232
234
236
238
240
EF1 EF2 EF3 EF4
Julia
n d
ate
Date of maximum flow
Natural
Present
The number of reversals is calculated by
dividing hydrological records into ‘rising’ and
‘falling’ periods in which daily changes in flows
are either positive or negative, respectively.
The annual average number of reversals
indicates whether a flow regime is influenced
only by precipitation input or includes anthropic
alterations. Figure 13 displays the mean annual
number of reversals at four EF sites, under
natural and present condit ions. In natural
conditions, the number of reversals decreases
downstream, reflecting the lower temporal
variability of the natural flow regime, mostly
FIGURE 12. Mean observed precipitation (1971-2005) in the subcatchments of four EF sites.
0
50
100
150
200
250
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35
Prec
ipit
atio
n d
epth
s (m
m)
10-day periods
EF1 EF2
EF3 EF4
caused by the mainstream’s integration of flow
fluctuations of the tributaries. Under present
conditions, the number of reversals at site
EF2 is greater than the number of reversals
recorded at site EF1. The increase in the flow
variability between sites EF1 and EF2 reflects
the great impact of dams located between the
two stations. As flows are still moderate in the
upper part of the basin, the operation of dams
can significantly alter the natural river regime,
unlike in the downstream areas where the
greater river discharge is less impacted by dam
operation.
22
FIGURE 13. Mean annual number of reversals recorded at four EF sites and at the basin outlet, under natural and present conditions.
0
20
40
60
80
100
120
EF1 EF2 EF3 EF4
Nu
mb
er o
f rev
ersa
ls/y
ear
Number of reversals
Natural
Present
Hydrological Changes Caused by ‘Climate Change’ Scenarios Under Present Water Infrastructure Development
Figure 14 displays the mean dry and wet season
flows at the four EF sites under present climate
conditions, and climate change projections from
the PRECIS RCM under A2 (PRECIS-A2) and
B2 (PRECIS-B2) scenarios. Upstream of site
EF3, dry and wet season flows produced by
PRECIS under the A2 scenario, precipitation is
lower than that in present climate conditions.
This tendency inverts downstream of the site
EF3, with higher dry and wet season flows at
site EF4 under PRECIS-A2 precipitation, in
comparison with flows produced by present
climate conditions. Similar patterns are observed
for precipitation (Figure 15), suggesting that
spatial variations in flow patterns originates
from the precipitation distribution over the
subcatchments of the four EF sites. Flows
produced by PRECIS-B2 precipitation are
systematically higher and lower than those
produced by PRECIS-A2 precipitation during
the dry season and wet season, respectively.
These flow differences most likely result from
the difference of PRECIS-A2 and PRECIS-B2
precipitation as displayed in Figure 15.
FIGURE 14. Mean dry and wet season daily discharge under present climate conditions (for the period 1971-2005), and for future
climate conditions under scenarios A2 and B2 (for the period 2071-2100).
0
50
100
150
200
250
300
350
400
450
EF1 EF2 EF3 EF4
Mea
n d
isch
arg
e (m
3 /s)
Dry season
0
500
1,000
1,500
2,000
2,500
3,000
3,500
4,000
EF1 EF2 EF3 EF4
Mea
n d
isch
arg
e (m
3 /s)
Wet season
Presentclimate
PRECIS-A2
PRECIS-B2
23
FIGURE 15. Mean dry and wet season precipitation under present climate conditions (observed precipitation averaged over period
1971-2005), and for future climate conditions under scenarios A2 and B2 (for the period 2071-2100).
0
100
200
300
400
500
600
700
EF1 EF2 EF3 EF4
Dry season
Present climate
PRECIS-A2
PRECIS-B2
0
200
400
600
800
1,000
1,200
1,400
1,600
1,800
EF1 EF2 EF3 EF4
Mea
n a
nn
ual
pre
cip
itat
ion
(mm
)
Mea
n a
nn
ual
pre
cip
itat
ion
(mm
)
Wet season
Figure 16 displays the average Julian dates
of minimum and maximum 1-day flow at four EF
site under present and future climate conditions.
The dates of minimum 1-day flow are highly
variable among stations and between different
climate scenarios. This variability is caused by the
operation of the dams, as the volume of controlled
outflow is of the same order of magnitude as
natural flow. As a result, a slight change in the
dam outflow results in a significant change in the
date of the minimum flow. The same phenomenon
explains the delay of the minimum flows under
scenarios A2 and B2. Julian dates of maximum
flows are delayed downstream as a result of
the delay in the onset of the monsoon in the
lower parts of the basin (Figure 12 and Figure
17). Maximum flows simulated from PRECIS-A2
precipitation occur later than maximum flows
simulated from PRECIS-B2 precipitation. This time
lag is caused by the slight difference in the timing
of the wettest period, which occurs earlier in the
case of PRECIS-B2.
FIGURE 16. Occurrence of annual extreme flow conditions.
0
20
40
60
80
100
120
140
EF1 EF2 EF3 EF4
Julia
n D
ate
Julia
n D
ate
Date of minimum flow
215
220
225
230
235
240
EF1 EF2 EF3 EF4
Date of maximum flow
Present
climate
PRECIS-A2
PRECIS-B2
24
Figure 18 displays the number of reversals
at four EF sites. At sites EF1, EF2 and EF3,
the number of reversals under climate change
conditions (PRECIS-A2 and PRECIS-B2) is lower
than the number of reversals observed under
present climate conditions. At site EF4, this
tendency reverses and the number of reversals
becomes greater under climate change conditions.
The greater number of reversals under climate
change conditions in the downstream part of
the catchment is caused by the higher variability
of precipitation. At each EF site, the number
of reversals is greater under PRECIS-B2 in
comparison with PRECIS-A2. This difference
results from the greater temporal variability of
PRECIS-B2 precipitation.
FIGURE 18. Mean annual number of reversals recorded at four EF sites, under present and future climate conditions.
FIGURE 17. Mean precipitation in the subcatchments of four EF sites, as predicted by PRECIS A2 and B2 scenarios. Averages
computed over the period 2071-2100.
0
20
40
60
80
100
120
140
160
180
200
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35
10-d
ay p
reci
pit
atio
n d
epth
(mm
)
10-d
ay p
reci
pit
atio
n d
epth
(mm
)
10-day period 10-day period
PRECIS-A2EF1
EF2
EF3
EF4
0
20
40
60
80
100
120
140
160
180
200
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35
PRECIS-B2EF1
EF2
EF3
EF4
0
20
40
60
80
100
120
EF1 EF2 EF3 EF4
Julia
n d
ate
Number of reversals
Present climate
PRECIS-A2
PRECIS-B2
25
Conclusions
This report explains the results of the first (known
to authors) attempt to analyze the impacts of
water infrastructure development in the entire
UGB, by comparing flow changes under natural
and present conditions. The analysis shows that,
on average, annual flows at present are 2-8%
lower than under naturalized conditions. Higher
flow reduction in the dry season (up to 70% in
February) is detected, compared to just a small
percentage change in the wet season. Therefore,
various dams and barrages constructed to date
have reduced mainly the flows during the dry
season – when irrigation water demands are
the highest. Flow regulation through dams and
barrages has also changed the timing of annual
extreme water conditions such as the date of
minimum and maximum flows. The change in
the timing of the minimum flow date is, however,
affected more than the maximum flows. Future
water resources development plans need to take
this into serious consideration in order to avoid
further detrimental impacts on the river ecology.
Also, the study simulated the impacts of
CC on water infrastructure development in
the UGB. The results suggest that both dry
and wet season flows under CC scenario A2
(scenario corresponding to high population
growth with slower per capita economic growth
and technological change) are lower than that
under present climate conditions at upstream
locations, but higher at downstream locations
and at the basin outlet. Flows simulated under
CC scenario B2 (corresponding to moderate
population growth and economic development
with less rapid and more diverse technological
change) are found to be higher during the dry
season, and lower during the wet season than
that under CC scenario A2. Under CC scenario
B2, the timing of the maximum flow period is
earlier than that under present conditions. This
basically means that the monsoon might start
earlier. Furthermore, greater temporal variability of
precipitation was found in the lower basin under
both the A2 and B2 scenarios. All these results
are very relevant to future water management
in the basin. In the upper parts of the basin,
especially in the Uttarkhand District, plans for
further hydropower development are underway.
Decrease in precipitation and flows will affect
water availability in the planned projects. Similarly,
change in the timing of the monsoon as well as
increased variability in precipitation, especially in
the lower parts of the basin, will affect the current
irrigation water regulation practices. Therefore,
some adjustment to agricultural practices such as
early sowing might be necessary if the projected
changes under the B2 scenario become a reality.
The present modeling study did not consider
scenarios of future water resources development
under future climate scenarios. As mentioned
above, in the UGB, especially in upper parts of
the basin, several hydropower dams are being
planned or operationalized. The impacts of the
already constructed Tehri Dam are now coming
into effect. The combined impact of future water
infrastructure development and climate on river
flow in the UGB and the availability of water for
agriculture and other uses, as well as the impacts
of CC on operation of infrastructure itself, is a
subject of a subsequent ongoing study.
It could be argued that precipitation scenarios
used to anticipate hydrological change under CC
are not reliable as they originate only from one
climate model: PRECIS RCM forced by HadCM3.
It is now accepted that climate models are not
able to accurately simulate precipitation, mostly
because of their inability to simulate actual climate
dynamics. For instance, Kingston et al. (2011)
showed that uncertainty in precipitation is the
main source of error in hydrological projections.
Even the use of averaged precipitations projected
by several climate models cannot reduce this
uncertainty as the variability between different
climate projections from different models is
high. However, Rupa Kumar et al. (2006), who
assessed the biases in PRECIS simulation over
India by comparing simulated and observed
precipitation, found that this RCM is able to
reasonably predict the climate over India, both in
terms of means and extremes. A second source of
26
uncertainty in the PRECIS precipitation projections
originates from the SRES scenarios, associated
with specific gas emission conditions that may
result in various precipitation conditions. In the
present study, two very contrasting scenarios
are used to cover the whole range of possible
precipitation changes. Consequently, even if the
CC projections used in this study are biased
because of the use of only one climate model, it
is moderated by the use of the two contrasting
SRES emission scenarios - A2 and B2.
The main constraint in this study (as well as
in all research carried out on water resources in
the Ganga Basin) is the availability of observed
data (on climate, hydrology, etc.). This limitation
severely affects model calibration and validation,
and, in the end - simulations of future scenarios.
The authori t ies responsible for observed
hydrological data management and sharing should
seriously consider opening their archives for water
research needs. Without more open policies on
observed data access, proper planning of water
resources development in the Himalayan parts
of India is impossible. The uncertainty in climate
projections is another major issue associated
with studies like this one. Improvement in CC
projections as well as access to more climate
data would certainly enhance the accuracy of the
simulations, and in the end – planning for the
future.
27
References
Ahmad, Q.K.; Ahmed, A.U.; Khan, H.R.; Rasheed, K.B.S. 2001. GBM regional water vision: Bangladesh
perspectives. In: Ganges-Brahmaputra-Meghna region: A framework for sustainable development, (eds.), Ahmad,
Q.K.; Biswas, A.K.; Rangachari, R.; Sainju, M.M. Dhaka: University Press Limited. pp. 31-78.
Akhtar, M.; Ahman, N.; Booij, M.J. 2009. Use of regional climate model simulations as input for hydrological models
for the Hindukush-Karakorum-Himalaya region. Hydrology and Earth System Sciences 13: 1075-1089.
Arnell, N.W. 2003. Relative effects of multi-decadal climatic variability and changes in the mean and variability of
�� �����0����!���������� �!8��0�0������������ ���� �� ���Journal of Hydrology 270(3-4): 195–213.
Arnold, J.G.; Srinivasan, R.; Muttiah, R.S.; Williams, J.R. 1998. Large area hydrologic modeling and assessment.
Part I: Model development. Journal of the American Water Resources Association 34(1): 73–89.
Benestad, R.E.; Hanssen-Bauer, I.; Chen, D. 2008. Empirical-statistical downscaling�� 9 �!�&��8� :����� 9� �� ?��
Publishing Co. pp. 300.
Bergström, S.; Carlsson, B.; Gardelin, M.; Lindström, G.; Pettersson, A.; Rummukainen, M. 2001. Climate change
impacts on runoff in Sweden-assessments by global climate models, dynamical downscaling and hydrological
modelling. Climate Research 16: 101-112.
������ #���@�B���C���#�2���$����D������! ���� �&�������� �������������'0������!�� �����F��� �G �����������#�
Bangladesh. Water International 36(3): 357-369.
��� H#� +�B�� �$$*�� J�&���� ��� �� ���� ����!� ��� � ��� ���� �!� ������� � ��� � ������ �&�� ��� ����� ����0� �����
Journal of Hydrology 303(1-4): 176-198.
Bouwer, L.M.; Aerts, J.C.J.H.; van de Coterlet, G.M.; van de Giesen, N.; Gieske, A.; Mannaerts, C. 2004. Evaluating
downscaling methods for preparing Global Circulation Model (GCM) Data for Hydrological Impact Modelling. In:
Climate Change in Contrasting River Basins, (eds.) Aerts, J.C.J.H.; Droogers, P. Wallingford, UK: CABI Publishing.
de Condappa, D. 2009. WEAP modeling for Indus and Ganges basin. Unpublished Report.
�����#�D�B�@�K ����#�L�F�� �$$N��O� �!� �! ����� �� ��������� ����� ��� � �0����� ���� ���� ���� �0�0�� � ��� ����� ��
northwest England. Climatic Change 80(3-4): 337-367.
Garg, K.K.; Bharati, L.; Gaur, A.; George, B.; Acharya, S.; Jella, K.; Narasimhan, B. 2011. Spatial mapping of
agricultural water productivity using the SWAT model in Upper Bhima Catchment, India. Irrigation and Drainage
3��� ���?����&0�� ������� �4�
Gosain, A.K.; Rao, S.; Basuray, D. 2006. Climate change impact assessment on hydrology of Indian river basins.
Current Science 90(3): 346-353.
Guo, S.; Wang, J.; Xiong, L.; Yin, A.; Li, D. 2002. A macro-scale and semi-distributed monthly water balance model
to predict climate change impacts in China. Journal of Hydrology 268(1-4): 1-15.
Gutowski, W.J.; Decker, S.G.; Donavon, R.A.; Pan, Z.; Arritt, R.W.; Takle, E.S. 2003. Temporal-spatial scales of
observed and simulated precipitation in central U.S. climate. Journal of Climate 16(22): 3841-3847.
Hay, L.E.; Clark, M.P. 2003. Use of statistically and dynamically downscaled atmospheric model output for hydrologic
simulations in three mountainous basins in the western United States. Journal of Hydrology 282(1-4): 56-75.
Hay, L.E.; Clark, M.P.; Wilby, R.L.; Gutowski, W.J.; Leavesley, G.H.; Pan, Z.; Arritt, R.W.; Takle, E.S. 2002. Use of
regional climate model output for hydrological simulations. Journal of Hydrometeorology 3(5): 571–590.
Jian, J.; Webster, P.J.; Hoyos, C.D. 2009. Large-scale controls on Ganges and Brahmaputra river discharge on
intraseasonal and seasonal time-scales. Quarterly Journal of the Royal Meteorological Society 135(639): 353-370.
IPCC (Intergovernmental Panel on Climate Change). 1996. Climate change 1995: Impacts, adaptation, and mitigation ������������� ��������������������������������������������������� � ��������������������������report of the Intergovernmental Panel on Climate Change, (eds.), Watson, R.T.; Zinyowera, M.C.; Moss, R.H.
Cambridge, UK: Cambridge University Press.
28
IPCC. 2000. IPCC special report on emission scenarios. Summary for policymakers. A special report of working
group III of the Intergovernmental Panel on Climate Change (IPCC. Geneva, Switzerland: Intergovernmental
Panel on Climate Change (IPCC). 20p.
Jha, M.K. 2011. Evaluating hydrologic response of an agricultural watershed for watershed analysis. Water 3(2):
604-617.
Jones, R.G.; Noguer, M.; Hassell, D.C.; Hudson, D.; Wilson, S.S.; Jenkins, G.J.; Mitchell, J.F.B. 2004. Generating high resolution climate change scenarios using PRECIS��+��P�?��D�����L���#��/��#�OK��Q$&����� �����
������&8RR&�� ������?�����R����R2G�LJ9�D������C�&���3����������P������Q#��$��4�
Kingston, D.G.; Thompson, J.R.; Kite, G. 2011. Uncertainty in climate change projections of discharge for the
Mekong River Basin. Hydrology of Earth System Sciences 15(5): 1459-1471.
Leander, R.; Buishand, T.A. 2007. Resampling of regional climate model output for the simulation of extreme river
������Journal of Hydrology 332: 487-496.
Liu, Y.B.; De Smedt, F. 2004. WetSpa Extension, ������������������� ������ Department of Hydrology and
Hydraulic Engineering, Vrije Universiteit Brussel, Belgium. 108p.
Mirza, M.M.Q. (ed.). 2004. The Ganges water diversion: Environmental effects and Implications. Dordrecht, the
Netherlands: Kluwer Academic Publishers.
Neitsch, S.L.; Arnold, J.G.; Kiniry, J.R.; Williams, J.R. 2005. Soil and Water Assessment Tool – Theoretical documentation (version 2005). Temple, Texas: Grassland, Soil and Water Research Laboratory, Agricultural
Research Service, Blackland Research Center, Texas Agricultural Experiment Station.
Pilling, C.G.; Jones, J.A.A. 2002. The impact of future climate change on seasonal discharge, hydrological processes
����/��������� �����O&&��:��/&� ��������������#�� �':�����Hydrological Processes 16(6): 1201-1213.
G������#�+�+���$$U�� J��!�����F��!����� ������!���8�L��� ���������&� ���� �! ���������&�����Water Policy 11(2): 168-190.
Rupa Kumar, K.; Sahai, A.K.; Krishna Kumar, K.; Patwardhan, S.K.; Mishra, P.K.; Revadekar, J.V.; Kamala, K.;
Pant, G.B. 2006. High-resolution climate change scenarios for India for the 21st century. Current Science 90(3):
334-345.
Seidel, K.; Martinec, J.; Baumgartner, M.F. 2000. Modelling runoff and impact of climate change in large Himalayan basins. International Conference on Integrated Water Resources Management (ICIWRM), December 19-21,
2000, Roorkee, India.
9 �!�#� 2�@� D�� ������#� O�K�@� K0���#� V�� �$$%�� +���� �!� ���� �� ��� ��� ��� � ������ ���&������ ��� ��������� ����
Gangotri Glacier basin, Himalayas. Journal of Geophysical Research 53(2): 309-322.
Singh, P.; Haritashya, U.K.; Kumar, N.; Singh, Y. 2006. Hydrological characteristics of the Gangotri Glacier, central
Himalayas, India. Journal of Hydrology 327(1-2): 55-67.
Srinivasan, R.; Ramanarayanan, T.S.; Arnold, J.G.; Bednarz, S.T. 1998. Large area hydrologic modeling and
assessment Part II: Model application. Journal of the American Water Resources Association 34(1): 91-101.
Verghese, B.G.; Iyer, R.R. 1993. Harnessing the eastern Himalayan rivers: Regional cooperation in South Asia.
New Delhi: Konark Publishers.
Wang, X.; Melesse, A.M. 2005. Evaluation of the SWAT Model’s snowmelt hydrology in a Northwestern Minnesota
watershed. Transactions of the American Society of Agricultural Engineers 48(4): 1359-1376.
Wilby, R.L.; Hay, L.E.; Leavesley, G.H. 1999. A comparison of downscaled and raw GCM output: Implications for
climate change scenarios in the San Juan River Basin, Colorado. Journal of Hydrology 225(1-2): 67-91.
::�� 3:����� : �� �0��� ���� V��0�4'J�� ��� �$������������� ��� �� ��������� ����� ���� O&&�� F��!�� ��� ���
New Delhi, India: World Wide Fund for Nature (WWF)-India. 21p. Available at ���!!���������� !����"���!������ "�������!��#�� " �� �$! (accessed on October 31, 2011).
Electronic copies of IWMI's publications are available for free.
Visit
www.iwmi.org/publications/index.aspx
IWMI Research Reports
142 The Impacts of Water Infrastructure and Climate Change on the Hydrology of the Upper Ganges River Basin. Luna Bharati, Guillaume Lacombe, Pabitra
Gurung, Priyantha Jayakody, Chu Thai Hoanh and Vladimir Smakhtin. 2011.
141 Low-Cost Options for Reducing Consumer Health Risks from Farm to Fork Where Crops are Irrigated with Polluted Water in West Africa. Philip Amoah,
��������������������� �����!���"�������������#�� �����������$�����%�
Konradsen. 2011.
140 An Assessment of Crop Water Productivity in the Indus and Ganges River Basins: Current Status and Scope for Improvement��&�����%�#��������������Sharma, Mir Abdul Matin, Devesh Sharma and Sarath Gunasinghe. 2010.
139 Shallow Groundwater in the Atankwidi Catchment of the White Volta Basin: Current Status and Future Sustainability. Boubacar Barry, BenonyKortatsi,
'�����$�����������������������'�������%�����(�������)���*������������
Joost van den Berg and Wolfram Laube. 2010.
138 Bailout with White Revolution or Sink Deeper? Groundwater Depletion and Impacts in the Moga District of Punjab, India. Upali A. Amarasinghe, Vladimir
+�������������������+����������(�������,����%�����-./.�
137 Wetlands, Agriculture and Poverty Reduction. Matthew McCartney, Lisa-Maria
������+������+��������+�������������+��0� ���+�� ���-./.�
136 Climate Change, Water and Agriculture in the Greater Mekong Subregion��������1���������'��������)�������#���2����3������ ����(�����!����!� �����
Vladimir Smakhtin, Diana Suhardiman, Kam Suan Pheng and Choo Poh Sze.
2010.
135 Impacts of Climate Change on Water Resources and Agriculture in Sri Lanka: A Review and Preliminary Vulnerability Mapping��(�������,����%�����4��������+���������)������#���������������������$��������-./.�
134 Evaluation of Current and Future Water Resources Development in the Lake Tana Basin, Ethiopia. Matthew McCartney, Tadesse Alemayehu, Abeyu Shiferaw
and Seleshi Bekele Awulachew. 2010.
133 Mapping Drought Patterns and Impacts: A Global Perspective�� (�������,����%�����4��������+������������(��������'���%��-..5�
Postal AddressP O Box 2075ColomboSri Lanka
Location127 Sunil MawathaPelawattaBattaramullaSri Lanka
Telephone+94-11-2880000
Fax+94-11-2786854
Websitewww.iwmi.org
I n t e r n a t i o n a lWater ManagementI n s t i t u t e ISBN 978-92-9090-744-2
ISSN 1026-0862
Related Publications
Bartlett, R.; Bharati, L.; Pant, D.; Hosterman, H.; McCornick, P. G. 2010. Climate change impacts and adaptation in Nepal. Colombo, Sri Lanka: International Water Management Institute (IWMI). 35p. (IWMI Working Paper 139)
www.iwmi.org/Publications/Working_Papers/working/WOR139.pdf
Cai, X.; Sharma, B. R.; Matin, M. A.; Sharma, D.; Gunasinghe, S. 2010. An assessment of crop water productivity in the Indus and Ganges River Basins: current status and scope for improvement. Colombo, Sri Lanka: International Water Management Institute (IWMI). 30p. (IWMI Research Report 140)
www.iwmi.org/Publications/IWMI_Research_Reports/PDF/PUB140/RR140.pdf
Jha, R.; Smakhtin, V. 2008. A review of methods of hydrological estimation at ungauged sites in India. Colombo, Sri Lanka: International Water Management Institute (IWMI). 24p. (IWMI Working Paper 130)
www.iwmi.org/Publications/Working_Papers/working/WOR130.pdf
Smakhtin, V.; Gamage, M. S. D. N.; Bharati, L. 2007. Hydrological and environmental issues of interbasin water transfers in India: A case of the Krishna River Basin. Colombo, Sri Lanka: International Water Management Institute (IWMI). 35p. (IWMI Research Report 120)
www.iwmi.org/Publications/IWMI_Research_Reports/PDF/PUB120/RR120.pdf
Venot, J. P.; Turral, H.; Samad, M.; Molle, F. 2007. Shifting waterscapes: Explaining basin closure in the Lower Krishna Basin, South India. Colombo, Sri Lanka: International Water Management Institute (IWMI). 60p. (IWMI Research Report 121)
www.iwmi.org/Publications/IWMI_Research_Reports/PDF/PUB121/RR121.pdf